ComfyUI > Nodes > TrentNodes > Point Tracker

ComfyUI Node: Point Tracker

Class Name

PointTracker

Category
Trent/LipSync
Author
TrentHunter82 (Account age: 0days)
Extension
TrentNodes
Latest Updated
2026-03-20
Github Stars
0.03K

How to Install TrentNodes

Install this extension via the ComfyUI Manager by searching for TrentNodes
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter TrentNodes in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

Point Tracker Description

Tracks a point in video frames using the Lucas-Kanade method for precise motion tracking.

Point Tracker:

The PointTracker node is designed to track a specific point across a sequence of video frames using the pyramidal Lucas-Kanade (LK) method. This technique is particularly effective for motion tracking in video processing, allowing you to follow the movement of a point with sub-pixel accuracy. The node is beneficial for applications such as lip-syncing in animation, where precise tracking of facial features is crucial. By leveraging image pyramids and iterative refinement, the PointTracker can handle large displacements and varying lighting conditions, making it a robust choice for dynamic video content. Its ability to generate tracking masks and previews further enhances its utility, providing visual feedback on the tracking process.

Point Tracker Input Parameters:

frames

This parameter represents the sequence of video frames in which the point will be tracked. It is a tensor containing the pixel data for each frame. The quality and resolution of these frames can significantly impact the accuracy of the tracking process.

start_x

The start_x parameter specifies the initial x-coordinate of the point to be tracked in the first frame. It is crucial for setting the starting position of the tracking process. The value should be within the frame's width.

start_y

Similar to start_x, the start_y parameter defines the initial y-coordinate of the point in the first frame. It determines the vertical starting position for tracking and should be within the frame's height.

window_size

This parameter defines the size of the window used for tracking the point. It must be an odd number and is clamped to the frame size. A larger window size can capture more context but may reduce precision. The default value is 31.

pyramid_levels

pyramid_levels indicates the number of levels in the image pyramid used for tracking. More levels can improve tracking robustness against large movements but may increase computational cost. The default is 4.

iterations

This parameter sets the number of iterations for the tracking algorithm at each pyramid level. More iterations can enhance accuracy but will also increase processing time. The default value is 10.

smoothing

The smoothing parameter controls the amount of temporal smoothing applied to the tracking results. A higher value can reduce noise but may also smooth out rapid movements. The default is 0.3.

search_radius_percent

This parameter defines the search radius as a percentage of the frame diagonal for template matching fallback. It helps in cases where the point moves significantly between frames. The default is 0.0, which uses the window size as the search radius.

Point Tracker Output Parameters:

points

The points output is a sequence of coordinates representing the tracked point's position in each frame. This output is crucial for understanding the movement path of the point across the video.

tracking_masks

tracking_masks are binary masks that highlight the tracked point's location in each frame. These masks provide a visual representation of the tracking process, aiding in verification and analysis.

preview

The preview output is an image tensor that shows the tracking results overlaid on the original frames. It serves as a visual confirmation of the tracking accuracy and can be used for further analysis or presentation.

Point Tracker Usage Tips:

  • Ensure that the initial point coordinates (start_x and start_y) are accurately set to improve tracking reliability.
  • Adjust the window_size and pyramid_levels based on the expected movement speed and complexity of the video content for optimal performance.
  • Use the smoothing parameter to balance between noise reduction and responsiveness to rapid movements.

Point Tracker Common Errors and Solutions:

"Invalid window size"

  • Explanation: The window size must be an odd number and within the frame dimensions.
  • Solution: Ensure the window_size is odd and does not exceed the frame's width or height minus two.

"Point out of frame bounds"

  • Explanation: The starting point coordinates are outside the frame dimensions.
  • Solution: Verify that start_x and start_y are within the valid range of the frame's width and height.

"Insufficient pyramid levels"

  • Explanation: The number of pyramid levels is too low for the given frame size or movement.
  • Solution: Increase the pyramid_levels to handle larger displacements effectively.

Point Tracker Related Nodes

Go back to the extension to check out more related nodes.
TrentNodes
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

Point Tracker